|Home | About | Journals | Submit | Contact Us | Français|
The widespread use of intravascular devices, such as central venous and hemodialysis catheters, in the past 2 decades has paralleled the increasing incidence of catheter-related bloodstream infections (CR-BSIs). Candida albicans is the fourth leading cause of hospital-associated BSIs. The propensity of C. albicans to form biofilms on these catheters has made these infections difficult to treat due to multiple factors, including increased resistance to antifungal agents. Thus, curing CR-BSIs caused by Candida species usually requires catheter removal in addition to systemic antifungal therapy. Alternatively, antimicrobial lock therapy has received significant interest and shown promise as a strategy to treat CR-BSIs due to Candida species. The existing in vitro, animal, and patient data for treatment of Candida-related CR-BSIs are reviewed. The most promising antifungal lock therapy (AfLT) strategies include use of amphotericin, ethanol, or echinocandins. Clinical trials are needed to further define the safety and efficacy of AfLT.
Antimicrobial lock therapy (ALT) utilizes prolonged instillation of a solution containing high concentrations of antimicrobial or antiseptic agents within an infected intravascular catheter, usually in conjunction with systemic antibiotics, in attempt to sterilize the catheter (1). Many catheter-related infections are related to intraluminal biofilms, which are notoriously difficult to treat, as biofilm-related (sessile) MICs (sMICs) are often dramatically elevated than MICs in planktonic (nonbiofilm) form. Thus, the basic approach of ALT is to utilize extremely high local concentrations of antimicrobial agents in the “lock” solution that are 100- to 1,000-fold higher than those used systemically. ALT has recently been recommended by current Infectious Disease Society of America (IDSA) practice guidelines as a first-line option for the management of catheter-related bloodstream infections (CR-BSIs) caused by coagulase-negative staphylococci and other select bacteria when catheter retention is indicated (2). However, for CR-BSIs caused by Candida species, current guidelines recommend catheter removal, particularly in nonneutropenic patients.
From a clinical standpoint, catheter reinsertion may be technically challenging and represents increased risk of morbidity and mortality for the patient. Patients may be critically ill, requiring intravenous access for medication administration, or total parenteral nutrition (TPN); they may be severely thrombocytopenic or coagulopathic or have extremely limited venous access. Depending on catheter location, potentially life-threatening complications such as pneumothorax, cardiac arrhythmias, arterial puncture, and hemorrhage can occur during insertion. Recent clinical epidemiology studies have reported that the rates of immediate complications for catheter insertions range from 7.5% in hematology patients to 7.7% in general hospital patients when performed by a dedicated central venous catheter (CVC) placement service using ultrasound guidance or 29.4% without ultrasound guidance (3, 4). The role of catheter exchange over a guidewire, which might be of benefit in some situations due to bacterial CR-BSIs (5), is undefined in CR-BSIs caused by Candida (6).
In addition to systematic efforts to reduce the incidence of CR-BSIs through a range of infection control approaches and new technologies, such as antibiotic-coated CVCs, clinicians have attempted antifungal lock therapy (AfLT) in situations where the need for catheter salvage in Candida-related CR-BSIs appears to outweigh the risks. Antimicrobial lock therapy has been utilized for CR-BSIs associated with totally implanted catheters (7, 8), which in principle should be amenable to antifungal lock therapy; however, clinical data are currently lacking. In this minireview, the in vitro, animal, and human data regarding the safety and efficacy of AfLT strategies are summarized.
Candida species are the fourth leading cause of health care-associated infections and the third most common cause of central line-associated bloodstream infections (9). Candida species are associated with the highest overall crude mortality of all nosocomial bloodstream infections, comparable to that of Pseudomonas and exceeding that of Staphylococcus aureus infections (10). Candida albicans is the most common fungal species associated with biofilm-related infections (11). In addition to device- and catheter-related infections, endocarditis and other prosthetic infections have also been associated with biofilm formation (12). In patients with candidemia, biofilm-producing strains of Candida species have been associated with increased morbidity and mortality compared to non-biofilm-producing strains (13).
C. albicans is a versatile commensal organism that possesses a number of attributes that enhances its ability to survive in diverse environments and transition from harmless commensal to invasive pathogen. Like many other microbial pathogens, C. albicans exists predominantly in biofilm (sessile) form, rather than living in freely floating planktonic form (14). Biofilms are complex, structured microbial communities that are attached to a mucosal or basal surface, encased in a matrix of exopolymeric substance (EPS) intermixed with host proteins (15). These sessile cells are highly organized and communicate via quorum-sensing molecules (16, 17). Whereas surface colonization is a commensal attribute, the abilities of C. albicans to adhere, form biofilms, and invade surfaces are virulence traits that are important for pathogenesis.
Candida biofilms are also characterized by increased resistance to conventional antifungal therapy, in particular amphotericin B and fluconazole (18). Mechanisms responsible for the intrinsic drug resistance of Candida biofilms include the following: limited drug penetration into the biofilm matrix, including binding of drug to EPS-related beta-glucan; reduced growth in the setting of limited nutrients and amino acid starvation; increased early expression of antifungal resistance genes, particularly those encoding azole efflux pumps; and the presence of a drug-resistant subpopulation of “persister” cells (19–21). Thus, increased antifungal resistance within fungal biofilms is a complex, pleiotropic phenotype due to multiple interacting mechanisms.
A wide variety of in vitro models have been utilized to study lock therapy for treatment and prevention of Candida biofilms. Many studies have utilized a static microplate model, usually with a silicone or polyvinyl chloride (PVC) disk to simulate commonly used catheter materials. Other models have used flow systems, such as a modified Robbins device (MRD), to study biofilms under conditions that are closer to physiological conditions. The ease of use of in vitro systems permits the study of a wide variety of candidate lock therapies to generate initial efficacy data. However, the limitations of these models include the static nature of the model, use of relatively high initial inocula of cells, and lack of cellular components that would reflect a host immune response. Since relatively few strains are often used in these studies, generalizability of the data may be limited. Nonetheless, these studies have generated valuable data regarding candidate lock therapies.
Amphotericin B deoxycholate (d-AmB) has reduced activity against Candida biofilms (22, 23). Lipid formulations have improved activity against Candida biofilms in vitro (24). Raad et al. compared EDTA, amphotericin B lipid complex (ABLC), and EDTA plus ABLC against five clinical isolates of Candida albicans and Candida parapsilosis (Table 1) (30). The combination of EDTA plus ABLC was the most effective against the biofilms, ABLC alone was moderately effective, and EDTA alone was not effective. Liposomal amphotericin (L-AMB) was studied in a static silicone catheter segment model against C. albicans, Candida glabrata, and C. parapsilosis biofilms. L-AMB (1 mg/ml) substantially, but not completely, inhibited biofilm metabolic activity. L-AMB (0.2 mg/ml) was the least effective, particularly with shorter lock times (<4 h) against mature, 5-day-old C. parapsilosis biofilms (32). The mechanism of enhanced activity of amphotericin B lipid formulations compared to d-AmB is unknown. It does not appear to be a unique property of the lipid vehicle, as lipid-associated nystatin (another polyene) did not inhibit Candida biofilms in vitro (24).
Azoles have poor activity against Candida biofilms. C. albicans has extremely high sMICs to fluconazole (FLC) in a standard static microplate biofilm model (33). This biofilm phenotype is particularly striking in C. albicans where the majority of clinical isolates remains fluconazole susceptible, as determined by Clinical and Laboratory Standards Institute (CLSI) antifungal susceptibility methods. Interestingly, in a silicone disk biofilm model, although simultaneous treatment with voriconazole (VRC) or posaconazole (PSC) and anidulafungin (AND) or caspofungin (CAS) produced indifferent antifungal activity, subinhibitory concentrations of echinocandins followed by PSC or VRC caused increased biofilm damage (34).
Echinocandins have excellent in vitro activity against Candida biofilms (35, 36). AND, CAS, and micafungin (MFG) demonstrated comparable, excellent antifungal activity against C. albicans (25) and against Candida tropicalis (26) in a static microplate model. Differences in echinocandin activity were mainly related to the occurrence of paradoxical growth at high doses, which was most prominent with CAS, followed by AND in these studies. Although this paradoxical effect is not thought to be clinically important for systemic therapy, it may be of theoretical concern if high doses of echinocandins are used in a lock strategy. Additional in vitro studies in a silicone catheter segment model have shown excellent activity of CAS and MFG against C. albicans biofilms (27) and superiority of CAS and MFG over PSC in the same model (28). Oncu et al. compared the activity of d-AmB, CAS, FLC, VRC, and itraconazole (ITC) at doses of 300×, 500×, and 1,000× the MICs against biofilms formed by one C. albicans strain and one C. parapsilosis strain in a silicone catheter segment model (29). Whereas d-AmB and CAS sterilized the segments by day 5, none of the azoles sterilized the segments. Overall, lipid preparations of amphotericin B and echinocandins demonstrated the greatest antifungal activity in these in vitro Candida biofilm models, whereas azoles demonstrated poor activity.
In vitro studies of combinations of standard antifungal therapy against Candida biofilms have shown various results, from indifference to synergism depending on the species and agents used (34, 37, 38). For example, in a static microplate model, the combination of d-AmB with PSC was synergistic, but d-AmB with CAS was indifferent against C. albicans clinical isolates (38). The role of combination standard antifungal therapy has not been specifically studied as antifungal lock therapy, and further studies of this approach are warranted.
A number of repurposed agents have been investigated for antifungal activity against Candida biofilms (Table 2). High-dose antibiotics, such as doxycycline and tigecycline, and alternative agents, including heparin and parabens, demonstrated substantial activity against C. albicans biofilms in a static microplate model (39–41). Other antibiotics with potential antifungal activity include minocycline with rifampin and ciprofloxacin with rifampin (42). Raad et al. examined the efficacy of several candidate lock solutions for antifungal activity against C. albicans biofilms in a modified Robbins device flow model (43). Of the solutions tested, minocycline plus EDTA, high-dose minocycline, and high-dose minocycline-EDTA all eradicated the biofilms. EDTA also enhanced the activity of ABLC in a silicone disk model of C. albicans and C. parapsilosis biofilms (30). Tetrasodium plus EDTA demonstrated antifungal activity against C. albicans in a biofilm flow model utilizing silicone catheter segments (47).
A wide variety of additional agents have been studied for in vitro activity against Candida biofilms. Extracellular DNA forms an important component of C. albicans biofilms (48); DNase, which selectively cleaves extracellular DNA, enhanced the activity of d-AmB, but not CAS or FLC, against Candida biofilms in a static microplate model (49). Catheters coated with chitosan, a biopolymer synthesized from crustacean-derived chitin, have antifungal activity in vitro and in an animal model of catheter-related biofilms (50). Taurolidine (2 H-1,2,4-thiadiazine-4,4′-methylenebis(tetrahydro-1,1,1′,1′-tetraoxide) is a substance with antiadherence and antimicrobial properties (51). A high-concentration taurolidine plus citrate catheter lock solution had some activity against C. albicans grown on silicone coupons in an MRD biofilm flow model (51). Other agents with in vitro activity include taurolidine-polyvinylpyrolidine (42) and aspirin and nonsteroidal anti-inflammatory agents (52). Further work is needed to determine whether any of these and other agents will be clinically useful as a component of a viable AfLT strategy.
Raad et al. compared minocycline, EDTA, and 25% ethanol (EtOH) separately and in various combinations in an MRD flow model and a silicone disk model (44). In the MRD model, the combination of EtOH-EDTA or EtOH-minocycline-EDTA resulted in complete elimination of C. parapsilosis biofilms when assayed at 24 h. In the silicone disk model, only the combination of EtOH-minocycline-EDTA resulted in complete elimination. Balestrino et al. compared 60% EtOH to 46.7% trisodium citrate (TSC) against biofilms formed on silicone catheter segments in a microfermentor (45). In this model, 60% EtOH eliminated biofilm colony growth after 20 min of incubation. In a static microplate model, 12.5% EtOH, N-acetylcysteine, EDTA, or talactoferrin each reduced C. albicans biofilm metabolic activity and CFU substantially, but only 12.5% EtOH reduced biofilm viability in a mixed biofilm with coagulase-negative staphylococcus, although sterilization was not achieved (46). In a systematic in vitro analysis of EtOH activity against C. albicans biofilms using a silicone disk model, incubation with 35% EtOH for at least 4 h eliminated biofilm viability and prevented colony regrowth (53). In a study of planktonic cells, Ghannoum et al. demonstrated that a solution of 5 mg/ml trimethoprim, 25% EtOH, and 3% EDTA (B-Lock solution) inhibited all Candida isolates tested, C. albicans, C. glabrata, C. krusei, C. parapsilosis, and C. tropicalis (25 isolates of each species) (54). Taken together, these in vitro studies suggest that EtOH has consistent activity against Candida species and appears promising as an AfLT strategy.
C. parapsilosis has been associated with catheter-related fungemias. Although there are some reports of reduced susceptibility of C. parapsilosis to systemic echinocandins (55, 56), this effect was not generally observed with in vitro studies that included C. parapsilosis biofilms. One in vitro study utilizing a silicone disk model of Candida biofilms found that mature (5-day-old) C. parapsilosis biofilms were inhibited to a lesser extent than C. albicans or C. glabrata biofilms by high-dose L-AMB, although no biofilms were eradicated (32). Although reduced susceptibilities may be a concern in C. parapsilosis CR-BSIs, given the favorable outcomes in clinical studies and the high doses of antifungal drug utilized in AfLT, this approach should still be feasible.
Polymicrobial bloodstream infections are a substantial clinical issue (57, 58); coagulase-negative staphylococci, Enterococcus, and Staphylococcus aureus are the most common copathogens in mixed Candida-bacterial BSIs (59). There is limited experimental data regarding ALT in this scenario. In a static microplate model using PVC disks, C. albicans enhanced the resistance of a non-slime-producing coagulase-negative staphylococcus to vancomycin. Staphylococcus epidermidis-C. albicans biofilms grown in an MRD flow model were resistant to d-AmB and FLC (60). In contrast to what has been observed clinically, S. aureus is poorly adherent to abiotic surfaces in vitro, but it forms a robust biofilm with C. albicans by associating with the dense network of underlying hyphae (61, 62). C. albicans enhanced S. aureus vancomycin resistance in a polymicrobial biofilm, whereas S. aureus did not alter C. albicans d-AmB susceptibility (61). In a silicone disk model of C. albicans-S. aureus biofilms, 30% EtOH was sufficient to prevent C. albicans regrowth, whereas 50% EtOH was required to prevent S. aureus regrowth within monomicrobial or polymicrobial biofilms (63). This is an area of current research interest (44, 54), which is needed to define the optimal approach for treatment or prevention of polymicrobial biofilms.
Several studies of AfLT on animals have been reported (Table 3). Mukherjee et al. studied the utility of ABLC as a lock strategy against C. albicans biofilms in a rabbit silicone catheter model (64). Catheters from all six animals in each treatment group were sterilized, compared to none of the catheters from animals in the control group. In an earlier study using this model, L-AMB was compared to FLC as a catheter lock strategy (65). All of the catheters from the six animals treated with L-AMB were sterilized, compared to only two of the six catheters from FLC-treated animals and none of the catheters from the animals in the control group.
In a different rabbit catheter model, Shuford et al. compared d-AmB to CAS in a 7-day lock model of C. albicans catheter infection with the same corresponding systemic antifungal therapy (66). Of the 16 animals in each arm, 16 catheters were sterilized in the CAS arm, 13 catheters were sterilized in the d-AmB arm, and none in the control arm. Lazzell et al. developed a mouse catheter model to examine the efficacy of CAS for the treatment and prevention of C. albicans biofilms (67). In the CAS lock arm (24-h dwell time), reduced fungal burden occurred in catheters and kidneys when used for either treatment or prevention of catheter-related infection compared to untreated controls. Overall, these results suggest that lipid preparations of amphotericin or an echinocandin may be promising for AfLT.
There are a limited number of case reports published describing the use of AfLT in various patient populations (Table 4). The majority of case reports describe AfLT experience in pediatric patients, ranging from infants to 18 years of age, but there are a few cases in adult patients. Among the case reports, the most commonly isolated fungi was C. albicans (9 of the 22 cases) followed by C. parapsilosis (4 of 22 cases). C. glabrata (2 cases), C. tropicalis (1 case), Candida guillermondii (1 case), and Candida lipolytica (1 case) were isolated less frequently.
The most commonly employed AfLT was d-AmB, with a combined catheter salvage rate of 76.9% (10 of 13 cases). Krzywda et al. reported using a significantly lower dose of d-AmB (0.33 mg/ml) compared to those used in other studies, which may have contributed to unsuccessful catheter salvage in all 5 episodes of fungemia in 2 patients (68). L-AmB AfLT was associated with a 60% (3 of 5 cases) salvage rate. In the only report of echinocandin lock therapy in the literature, CAS (3.33 mg/ml) combined with systemic CAS for 14 days was used to successfully treat C. lipolytica CR-BSIs (69). Blackwood et al. reported the successful use of a 70% EtOH lock solution for catheter salvage in three pediatric patients with invasive candidiasis (70).
AfLT was most commonly employed for 14 days (using various instillation times) after the last negative blood culture in conjunction with systemic antifungal therapy. A wide variety of catheter types are described and do not appear to play a role in directing the type of AfLT that was utilized or duration of therapy.
There were a few patients with multiple recurrences of candidemia, requiring multiple attempts at treatment and catheter salvage. In some cases, the patient was apparently successfully treated the first time but developed a recurrent infection, necessitating catheter removal. Regardless of the type of AfLT used and systemic antifungal therapy, the overall clinical catheter salvage rate reported in the literature is 77%. Although the outcomes presented in these case reports are encouraging, these data need to be interpreted with caution, due to the effects of publication bias and lack of a true “denominator” of cases.
On the basis of the available in vitro and in vivo data, echinocandins, amphotericin (L-AMB or d-AmB), and EtOH appear to be the most promising AfLT strategies. Limitations of EtOH lock therapy center mainly on the potential incompatibility with polyurethane catheters, which are commonly used for central venous access. Side effects such as flushing and lightheadedness can be avoided by aspirating the EtOH lock prior to catheter usage, rather than flushing it through the line. The roles of adjunctive and repurposed agents as a component of AfLT remain to be defined. However, there is little comparative data between these potential AfLT strategies and insufficient clinical data to permit specific recommendations for the use of AfLT in the management of CR-BSIs caused by Candida spp. at this time. The optimal systemic antifungal therapy to be used in a situation of catheter salvage is not well-defined, but on the basis of the available evidence, we would recommend an echinocandin, with liposomal amphotericin as a suitable alternative, over an azole. Randomized clinical trials of the most promising AfLT and systemic therapy combinations are needed to definitively evaluate the safety and efficacy of AfLT.
This work was supported in part by grants from the Department of Veterans' Affairs (to S.A.L.), and the Biomedical Research Institute of New Mexico (to S.A.L.).
We declare that we have no conflicts of interest.
Published ahead of print 15 October 2012